FRISP Workshop 2010

Abstracts

Ocean properties near and below the Fimbul Ice Shelf

Ole A. Nøst

The Fimbul Ice Shelf in the Weddell Sea reaches above the Antarctic Continental slope to ocean depth below 2000m. Only a few data records exists from the fringes below the ice shelf, but recently the first hot water drilled access holes through the ice shelf were accomplished. The first through 220 m ice close to the sill (M1), the second well inside the Jutul Basin with 395m ice thickness (M2), and the last through 191m of ice in one of the eastern connections (M3). We present vertical profiles of salinity and temperature from the three locations, as well as the first ~14 days of current meter data from M1.

The profiles reveal a water column at the freezing point in the layer directly below the ice shelf of 20-50 m thickness. At M1, a shallow layer with in situ supercooling indicating freezing. The supercooling disappeared after a few days. Most of the water properties generally remain along one single Gade-line in TS space, i.e. indicating that there is mainly one water mass in contact with the ice. Relatively warm water (T~-1.6 °C) is found near bottom (850m) at M2, and two intermediate layers are evident at mid depth (M3).

Most models allow direct access of some Warm Deep Water across the 650 m deep sill. This is, however, not in agreement with our observations, where water is only slightly warmer than the surface freezing point. Inflow of warmer water therefore seems limited. Pathways of inflow beneath the Fimbul ice shelf remains unclear, but the 6 deployed current meters will document annual cycles and mean currents when they are re-visited next year.

Ice Shelf Water export from the Filchner-Ronne Ice Shelf

Svein Østerhus¹ and Kjersti Strand²

After the discovery and mapping of the Ice Shelf Water (ISW) overflow over the Filchner sill in the Weddell Sea, instrumented moorings "S2" were deployed during the Norwegian Antarctic Expedition 1977/78. The position for the mooring S2 proved to be a key site for monitoring the ISW overflow and is selected to be a part of the global net of monitoring sites under CLIVAR (www.clivar.org) and OceanSITES (www.oceansites.org). Since the establishment of the S2 mooring, a number of yearlong time series have been obtained. Since 2003 the mooring has been serviced on a biannual basis by the BAS ships supplying the Halley station. During the IPY project BIAC (www.bccr.no/biac) we have designed and deployed a new observatory for long term monitoring of the export of ISW. The observatory is equipped with high quality sensors for water velocity, temperature, salinity, and dissolved oxygen and at a later stage we will add sensors to measure sea ice thickness. The data are downloaded by means of an acoustic link to a surface unit. Here we will present some results and discuss design for future observatories.

Melting beneath Larsen C Ice Shelf - a 'hydrographic' estimate

Hartmut H. Hellmer¹, O. Huhn², D. Gomis³, R. Timmermann¹ and M. Schröder¹

For the last five decades the Antarctic Peninsula faces the strongest atmospheric warming on Earth with severe consequences for its glaciers, ice shelves, sea ice cover, and the surrounding marginal seas. We analyzed hydrographic data from the northwestern Weddell Sea continental shelf of three austral winters (1989, 1997, and 2006) and two summers following the last winter cruise. The whole water column north of 66° S freshened by ~0.1 between the winters of the 17-year period, replaced by deep waters of the Central Bransfield Strait Basin every summer. The discussion of the causes for the salinity decrease, supported by tracer analysis, favors the increased input of glacial melt from underneath Larsen C Ice Shelf. However, the 2-m/yr melt rate, necessary for a year-by-year freshening, could be reduced to 0.38 - 1.33 m/yr due to a recent (modeled) increase of precipitation and a retreat of the sea ice cover in the northwestern Weddell Sea.

Changes of CDW on the Amundsen Sea Shelf as a major cause for ice sheet melt

M. Schröder¹, H.H. Hellmer¹, A. Wisotzki¹ and S.S. Jacobs²

The joint geophysical and oceanographic German Expedition with RV Polarstern into the Amundsen Sea in January to April 2010 was an important contribution to the long-term monitoring of CDW characteristics on the Amundsen Sea continental shelf, continuously conducted by colleagues in the US (LDEO) and UK (BAS), as part of the international ASEP (Amundsen Sea Embayment Project). It provided additional data for the validation of a coupled ice-ocean finite element model (FESOM) applied to the Southern Ocean as part of AWI's research programme PACES.

The oceanographic part of the cruise had three main objectives:

  • Determination of CDW routes and properties from the continental slope to the ice shelf fronts
  • Measurement of other shelf water characteristics to detect freshwater changes due to basal melting
  • Extension of collaborations with international colleagues investigating the ocean’s role in WAIS thinning

For the Amundsen Sea it is entirely plausible that ocean influence on the WAIS (West-Antarctic Ice Sheet) could increase from changes in ocean temperature, heat transport and vertical thermohaline structure, in response to altered atmospheric forcing, sea ice production, and ice shelf morphology. The first oceanographic stations occupied on the Amundsen continental shelf in 1994 have since been supplemented by late summer measurements in 2000, 2003, 2006, 2007, 2008, and 2009. This work has revealed that the 'warm', salty CDW gains access to the continental shelf near the sea floor, particularly in the eastern sector, and ponds in glacially scoured troughs that extend deep beneath the ice shelves. Where the ice shelves have deep grounding zones, they can be exposed to seawater more than 3°C above the in situ melting point. This drives basal ice shelf melting rates orders of magnitude faster than beneath the largest ice shelves. Substantial thermohaline variability is apparent in some of the repeated late summer observations, but little is yet known about the seasonal cycle or interannual variability. Also lacking is information about the strength of the ocean circulation, heat transport to the ice shelf caverns, and how efficiently that heat is used for basal melting. Ocean warming documented at distant locations like the Antarctic Circumpolar Current (ACC) may be transmitted directly to the exposed glacial ice, or weakened in mixing across the Antarctic Slope Front. Heat transport from the continental shelf break to the ice shelf caverns may also be influenced by mixing over the rough bottom topography, tidal currents, winds, sea ice production, icebergs, and meltwater impacts on the pycnocline.

The talk will give first results from the just finished cruise.

Observations beneath Pine Island Glacier West Antarctica

Pierre Dutrieux

Thinning ice in West Antarctica, resulting from acceleration in the flow of outlet glaciers, is currently contributing about 10% of the observed rise in global sea level. Pine Island Glacier (PIG) in particular has shown nearly continuous acceleration and thinning throughout the short observational record. Rapid thinning of the floating ice shelf that forms where the glacier reaches the coast, driven by changes in ocean heat transport beneath it, and the consequent inland retreat of the line separating grounded and floating ice has been postulated to be the cause. Here we report evidence gathered by an Autonomous Underwater Vehicle (AUV) operating beneath the ice shelf, that PIG was recently grounded on a transverse ridge in the sea floor. Warm seawater now flows through a widening gap above the submarine ridge, rapidly melting the thick ice of the newly-formed upstream half of the ice shelf. The present evolution of PIG is thus part of a longer-term trend that has moved the downstream limit of grounded ice inland by 30 km into water that is 300 m deeper than over the ridge crest. The ocean circulation beneath the glacier tongue is fairly vigorous and topographically controlled, reaching 20 cm/s along the ridge crest and near the grounding line.

Modelling heat transport across the Antarctic Slope Front - Residual vs. Eulerian overturning

Tore Hattermann

Ice shelves at the coast of Dronning Maud Land are situated right at the continental slope. In this area, basal melting is mainly determined by the amount of Warm Deep Water (WDW), entering the ice shelf cavity across the Antarctic Slope Front (ASF). Here we use a high resolution primitive equation model with terrain following coordinates to simulate the circulation above the continental slope, and underneath the ice shelf. The model's simplified geometry is based on the situation at the Fimbul ice shelf. The simulated temperature profile in the ice shelf cavity corresponds surprisingly well with observations below the Fimbul ice shelf. We find a residual overturning based on mesoscale eddies to be the reason for mixing of dense WDW across the ASF and onto the continental shelf. This is counteracted by an Ekman-induced Eulerian overturning, caused by the preliminary easterly winds near the shore.

Our results show that the eddy overturning across the ASF is crucial for the exchange of the ice shelf cavity with the open ocean and needs to be taken into account while estimating basal melt rates.

Oscillating dense currents

Paul Holland

The flow of dense polar shelf waters down continental slopes is a critical component of the global ocean circulation, and an analogous flow of buoyant melt water occurs up the base of ice shelves. Observations from the Ross Sea suggest that such plumes can be heavily impacted by tidal variability, and many of the world’s important dense water sources are located in tidally-active areas. Tides affect the source of dense water (by modulating the location of hydrographic gradients) and control the subsequent plume mixing and flow path. In an effort to separate these effects, dense plumes are modeled here by extending a classical one-dimensional plume model to two unsteady scenarios in which the plume path is fixed. The first case features a pulsed release of dense water into a stagnant ambient, and the model predicts gravity waves that propagate down the plume. The second case is of a steady-sourced plume flowing through an ambient that has uniformly-oscillating flow. This drives fluctuating shear between plume and ambient that leads to variable entrainment of ambient fluid into the plume, leading to standing ‘entrainment waves’. These different effects are investigated and their impact upon the interpretation of ocean observations considered.

Fimbulisen I - Radar2: Airborne radio echo sounding and radar satellite imagery revealing the thickness and internal structure of the ice shelf

Daniel Steinhage¹ and Angelika Humbert²

Fimbulisen, an ice shelf situated at the coast of Dronning Maud Land, East Antarctica, has a size of about 37.000 km2. Its thickest part, the extension of the fast flowing Jutulstraumen, reaches an ice thickness of up to 985 m, while west and east of the ice stream the ice is moving slower and is on average about 300 m thick.

This study uses two sources of radar data to identify the three dimensional geometry of the ice shelf, as well as the internal structure. Radar imagery acquired by TerraSAR-X is used for a classification based on surface features and airborne radio echo sounding data, surveyed by the Alfred Wegener Institute, reveal the internal structure of the ice shelf and its thickness. A focus of the study is the highly rifted western part of the ice shelf.

The airborne radio echo sounding profiles were collected between 1996 and 2008 with a frequency of 150 MHz and pulse length of 60 ns, respectively 600 ns. The data evaluation comprises the determination of the ice thickness as well as the extraction of changes and features of the internal structure of the ice in parts of the rift zone. High-resolution radar imagery acquired by the TerraSAR-X in 2008 and 2010 is used to evaluate principal deformation axis at characteristic locations, to detect crack modes as well as to classify zones of similar structural characteristics. Both data sets were combined in order to determine regions of similar mechanical properties of Fimbulisen for the numerical modelling study presented in Fimbulisen II - modelling.

Seismic methods on the ice shelf: A comparison between vibroseis and explosive seismics on Ekströmisen

C.M. Hofstede¹, O. Eisen¹, Y. Kristoffersen², A. Lambrecht³, C. Mayer⁴, L. Kindermann¹ and R. Blenkner²

We present first-time-ever results of active seismic measurements on an ice shelf with a vibroseismic source. The measurements were conducted in the 2009/10 field season in Dronning Maud Land, Antarctica. A Failing Y-1100 on skis with a mass of 16 t was used on the Ekströmisen ice shelf where ice is about 100-200 m thick and overlies about the same amount of water. The goal was to investigate the feasibility of vibroseismic operations on a porous firn layer to image the internal structure of the ice and the underlying sediments as well as ice and water column thickness. In comparison with conventional explosive seismics, where explosives are detonated in 20-m deep boreholes, vibroseimics does not require any drilling and can thus be considered a true surface measurement. In combination with a snow streamer the vibroseismic operation would enable long seismic traverses in comparably short time periods. Vibroseismics seems a promising site survey method for core drilling campaigns like the ANDRILL project but also to map the nature of the ice bed contact over long traverses. In addition to a comparison with conventional explosive shots we also recorded the signals at the PALAOA observatory with two calibrated hydrophones underneath the ice shelf, which enables the determination of absolute amplitudes and thus absorption within the ice.

Observation and parameterisation of ablation at the base of Ronne Ice Shelf, Antarctica

Adrian Jenkins, Keith W. Nicholls and Hugh F.J. Corr

Parameterisations of turbulent transfer through the oceanic boundary layer beneath an ice shelf are tested using direct measurements of basal ablation. Observations were made in the south-western part of Ronne Ice Shelf, about 500 km from open water. The mean basal ablation rate was measured over a month-long and a year-long period using phase-sensitive radar to record the thinning of the ice shelf. Ocean temperatures were observed within about 25 m of the ice shelf base over the period of the radar observations, while the tidally-dominated ocean currents were estimated from tidal analysis of co-located current observations from an earlier period. Ablation rates derived using these ocean data and a number of bulk parameterisations of turbulent transfer within the boundary layer are compared with the direct measurements. The ablation rates derived using a parameterisation that explicitly includes the impact of ocean currents on the turbulent transfer of heat and salt match the observations to within 40%, and with suitable tuning of the drag coefficient the mismatch can be reduced below the level of the observational errors. Equally good agreement can be obtained with two slightly simpler, current-dependent parameterisations that use constant turbulent transfer coefficients, and the optimal values for the coefficients at this particular location on Ronne Ice Shelf are given.

Marine ice within the Larsen C Ice Shelf and its implication for stress distribution

Daniela Jansen

Many ice shelves on the Antarctic Peninsula have retreated in the light of ongoing climatic warming. The dramatic break-ups of the Larsen A and B ice shelves in 1995 and 2002 have seen particularly pronounced scientific and public interest. As climatic warming is progressing southwards on the Antarctic Peninsula, scientific focus has been shifting to the southern neighbour, the Larsen C ice shelf. This ice shelf is one of the largest in Antarctica and is buttressing a considerable number of outlet glaciers that evacuate large quantities of ice from the Antarctic interior. Identifying the role of internal and external control mechanisms in regulating the present stability of the Larsen C ice shelf is therefore a global research priority.

The fundamental hypothesis of the SOLIS (Stability Of Larsen Ice Shelf) project, funded by the UK Natural Environment Research Council, is that mechanically soft ‘flow stripes’ sandwiched between mechanically stiffer units of glacier ice represent a governing control on ice shelf stability. Satellite and structural glaciological observations suggest that the softer flow stripes critically control rates of rift propagation. Within the SOLIS project these flow stripes are investigated by a combination of remote sensing, geophysical measurements and modelling. Here we present initial geophysical results from the first and second field season of the SOLIS project as well as initial model studies.

In the first field season anisotropic, multi-component seismic reflection and ground-penetrating radar (GPR) surveys were conducted to determine the detailed structure of the ice shelf at two control sites within and on either side of an inferred softer flow stripe on the southern Larsen C ice shelf, and to elucidate the ice mechanical properties at these sites. An anomalous reflector in about 100 m depth in the area of the flow stripe indicates that the different response to stresses within these flow stripes might be due to the presence of marine ice. The slightly diffuse reflector might be interpreted as the base of a meteoric ice body, while the seismic measurements yield an ice shelf thickness of about 320 m, indicating the presence of marine ice in the lower ice column. In the second field season a 4 km GPR grid was measured located on the same flow stripe further upstream, with the aim to get a three dimensional view of the boundary between meteoric and marine ice. The reflector is clearly visible within all profiles crossing the flow stripe and is dipping towards flow direction, reflecting accumulation at the ice shelf surface. No seismic survey was conducted in the second field season, but GPS elevation data indicates that marine ice must be present in the lower part of the ice column to reach hydrostatical equilibrium.

Results from both field seasons have been fed into a continuum-mechanical ice shelf model to investigate the impact of the marine flow bands on the stress regime of the Larsen C Ice Shelf. Model results from a very simple first approach by implementing the flow stripes via ice temperature show, that marine ice-rich, relatively warm flow bands in the north and south of Larsen C Ice Shelf promote an acceleration of the central part of the ice shelf in comparison to results without the implementation of flow stripes. An interesting question yet to be answered is where and under which conditions the presence of soft flow stripes represents a stabiliser (preventing rift growth) or possible instability (decoupling between different flow units, acceleration) of the ice shelf.

Annual to sub-hourly scale movement of the Mertz ice-tongue. Interaction with the ocean

L. Lescarmontier^1,2, B. Legrésy¹, R. Coleman², C. Mayet¹ and L. Testut¹

In the IPY CRACICE project, we are investigating the processes involved in the development of rifts across the Mertz Glacier Tongue leading up to its calving in February 2010. The length of the floating section of the Mertz Glacier was about 140km, with about 100km protruding out from the coastline in East. In November 2007, we deployed a network of GPS beacons on the glacier, and collected two months of data from two sites (GPS4&5), around the main rift on the southeastern side of the glacier. At this time, the Mertz glacier revealed active rifting through two kinds of oceanic influences: daily and sub-daily ones.
Between the 12th and the 13th, the B9B iceberg collided with the Mertz ice tongue causing a calving event of an 80km by 30km iceberg.

We used these GPS data to provide a detail view of motions at a fine time scale, and other velocities derived from ERS-INSAR, and pairs of SAR, or Landsat images, together with a DEM derived from stereo SPOT images and ice thickness data from Radio Echo sounding.

From a first processing of the GPS data, we are able to show that it is freely floating ice tongue with a horizontal velocity of about 3m/day toward the north-east. A kind of stick-slip effect appears to occur at daily scales. We see a modulation of the velocity at time scale of around 14 days. However, we observe that the maximum speed occurs a few days after spring tides. When the rifts on both east and west sides of the glacier were already well-developed, we find that the downstream and upstream parts of the ice-tongue were hinging horizontally at the rift on a daily to annual time scale.

With an accurate GPS processing, using Gins-pc and Precise Point Positioning processing (PPP), we are able to look at the sub-hourly scale signals. Using spectral analysis we identify three main energetic vibration modes around 6-20 minutes, 1.5-3.5 hours, and from 6.6 to 15 hours. We then compare the observed GPS vibrations with calculations of the fundamental mode of vibrations of a beam using different geometry cases.

Given that GPS4 & GPS5 were situated on each side of the rift, we investigate the impact of the vibration on the opening of the rift by a comparison of their signals. Then, we look at the possible resonance with an ocean forcing and their impact on calving processes.

Finally, we focus on the effect of atmospheric forcing on the ocean through TUGO model and its importance on local ocean level.

Calving of Mertz Glacier

Neal Young^1,2, Benoit Legresy³ and Richard Coleman^1,4

In mid-February 2010, iceberg B09B collided with the Mertz Glacier tongue. The collision precipitated the calving of massive iceberg, C28, measuring 78 km long and between 33 and 39 km wide. This calving event removed about 80% of the tongue, which had protruded about 100 km from the Antarctic coastline. The calving had been anticipated, as rifts cutting across the tongue had been developing over many years, but the timing and collision was not.

Two major rifts had been developing from opposite sides of the tongue over previous years. In the 1990s a large rift propagated in towards the centre-line from the eastern side. At that time the tongue exhibited some horizontal flexing, with tidal currents. In early 2000s, the north-west corner of the tongue appeared to be in contact with the sea bed. A second rift started to propagate in from the western side, and the fracture was almost complete. The rifts exhibited some opening and closing, but the tongue otherwise continued to extend in length.

Iceberg B09B had been grounded for many years to the north-east of the Mertz Glacier. In late November 2009 it started to move slightly. Over the next few weeks it rotated about a point near the eastern end, like a large pendulum, until its western end eventually struck the glacier tongue on about 07-February.

A composite series of satellite images show the daily behavior of the iceberg(s) and glacier tongue. An animation generated from this series illustrates the motion of the various elements. The progress of B09B is not smooth. There appears to be a shorter term, backwards and forwards, motion probably linked with tidal currents superimposed on the general trend towards the tongue. It also appears that there are multiple collisions involved. Iceberg C28 did not separate at the first impact, but about a week later. The calving front comprised some length of the existing rifts plus an additional segment where the ice sheared through to produce a clean front. C28 drifted off to the west and B09B now remains grounded on its pinning point to the north east of the remainder of the glacier tongue. C28 later fractured into two large and some smaller sections upon another collision with a high point on the sea bed to the west. GPS systems had been deployed on the glacier either side of the rift early in the 2009-2010 season. When the GPS data are recovered, detail information on the sequence of events will be revealed.

Separation of icebergs and sea ice in SAR images

Christine Wesche

The largest loss term in Antarctic mass balance is iceberg calving from the ice shelves and to estimate the amount of the loss, it is necessary to observe icebergs in every size. Because current mass loss calculations only include icebergs with an edge length of > 10 km, we focus on smaller icebergs (0.1 to 10 km edge length) in a test region north of Berkner Island in the Weddell Sea. Images of the ENVISAT ASAR at different imaging modes are used to analyse the backscattering coefficients of icebergs depending on the season. To detect icebergs in SAR images we need to find differences to the surrounding sea ice. Therefore, the backscattering coefficients of the sea ice are analysed for seasonal variations as well.

Statistical analyses of the backscattering of icebergs and sea ice in ASAR image mode data show varying backscatter coefficients over the period of one year. The radar intensity contrast between icebergs and sea ice is smallest in the summer months and highest in winter and spring.

The iceberg and sea-ice backscattering is investigated for seasonality in medium and low resolution ASAR images as well and compared to the results derived from image mode data. We will also include other frequency bands from other sensors to achieve a complete view of iceberg signature in radar images and their contrast to the surrounding. These statistics will improve the automatic extraction of icebergs from SAR images. As a next step, the extracted iceberg positions will be used to calculate the drift.

Drift and breakup of large icebergs

Neal Young

Over the past several years a number of very large icebergs have drifted away from the Antarctic, across the Southern Ocean into the south Indian Ocean, and generally eastwards to areas south of Australia and New Zealand. These icebergs all have their origins in the major calving events that occurred along the front of the Ross Ice Shelf. They followed different paths up to mid-latitudes and were exposed to open ocean conditions for different periods.

On the other hand they each exhibit similar behaviour in their pattern of breakup. All of them eventually disintegrated into many small sections. Prior to that, small sections were intermittently lost from the edges of the large icebergs, and the main iceberg(s) sometimes split into two still large sections.

For the case of iceberg B17B, the berg was imaged using several different systems (Landsat, SAR, MODIS). There does not appear to be evidence of significant melt on the surface of the berg, and no substantial melt ponding was observed. The key feature of the progression of the icebergs towards their ultimate disintegration is progressive thinning by basal melting. If surface melt does play a role then it is in the final act rather than a driver of the breakup.

Interactive Icebergs in a Coupled Global Circulation Model

Torge Martin

Icebergs are an important part of the fresh-water cycle and, until now, have not been explicitly represented in Intergovernmental Panel on Climate Change (IPCC) class coupled global circulation models (CGCMs) of the climate system. In this study we examine the impact of introducing interactive icebergs in a next-generation CGCM designed for 21st Century climate predictions. The frozen fresh-water discharge from land is used as calving to create icebergs in the coupled system which are then free to evolve and interact with the sea-ice and ocean components. Icebergs are fully prognostic, represented as point particles and evolve according to momentum and mass balance equations. About 100,000 individual particles are present at any time in the simulations but represent many more icebergs through a clustering approach. The various finite sizes of icebergs, which are prescribed by a statistical distribution at the calving points, lead to a finite life-time of icebergs ranging from weeks, for the smallest icebergs (60 m length), up to years for the largest (2.2 km length). The resulting melt water distribution seen by the ocean enhances deep-water formation, in particular on the continental shelves, relative to the model without icebergs.

Modeling Ice Shelf Cavities in a z-Coordinate Ocean General Circulation Model

Martin Losch

Processes at the ice shelf-ocean interface and in particular in ice shelf cavities around Antarctica have an observable effect on the solutions of basin scale to global coupled ice-ocean models. Despite this, these processes are not routinely represented in global ocean and climate models. It is shown that a new ice shelf cavity model for z-coordinate models can reproduce results from an intercomparison project of earlier approaches with vertical sigma- or isopycnic coordinates. As a proof of concept, ice shelves are incorporated in a 100 year global integration of a z-coordinate model. In this simulation, glacial melt water can be traced as far as north as 15°S. The observed effects of processes in the ice shelf cavities agree with previous results from a sigma-coordinate model, notably the increase in sea ice thickness. However, melt rates are overestimated probably because the parameterization of basal melting does not suit the low resolution of this configuration.

Ice shelf freshwater flux estimates

Michael P. Schodlok

Dense water masses that form on the continental shelves around Antarctica and spread into the global abyss are a major contributor to the global overturning circulation. Until recently the Estimating the Circulation and Climate of the Ocean, Phase II (ECCO2) project did not adequately represent many high latitude processes, including ice shelf ocean interactions. This work concerns the addition of an ice shelf package to the ECCO2 ocean state estimates. The ECCO2 solutions are obtained by fitting a high resolution (18 km horizontal grid spacing and 50 vertical levels) global-ocean and sea-ice configuration of the Massachusetts Institute of Technology general circulation model (MITgcm) to the available ocean and sea-ice data. Heat and freshwater exchanges at the ice shelf-ocean interface is based on the ISOMIP parameterisations. Ice, Cloud, and land Elevation Satellite (IceSAT) data provided the ice shelf thickness and Antarctic Bedrock Mapping Project (BEDMAP) provided the water column thickness in the cavities. Ice shelves were treated as floating slabs on water without flexural deformation. The results presented here were taken from a Southern Ocean subdomain of the global model. Freshwater Flux estimates for the period 1979 to 2008 amount to 59 mSv +- 7mSv. Whilst freshwater fluxes from various ice shelves are presented here for the first time, the major ice shelves show melt rates comparable to estimates from regional studies, e.g. BRIOS. Time series of the freshwater fluxes reveal no regional preference for increased/decreased freshwater input. While Moscow University IS shows a decrease its neighbour Shackleton IS has a stable freshwater flux. As the additional freshwater input stabilizes the water column and leads to increased sea ice thicknesses, this new package is used to further improve the global model solution.

Impacts of ice sheet mass balance changes on the global ocean

Rüdiger Gerdes

Changes in the mass balance of ice sheets influence the oceanic mass distribution and thus sea level. Beyond the direct effects of the fresh water input, indirect effects through the impact on the oceanic stratification, mixing and circulation have to be considered. Regionally, these changes can be most important. To investigate the consequences of rapid melting of ice sheets, water hosing experiments have been performed with a group of diverse climate models. The results have contributed to the Coupled Model Intercomparison Project (CMIP). These experiments were more geared to elucidate changes at the end of major glaciations and can not easily be employed to current changes and anthropogenic trends. A more suitable experiment for this purpose has been suggested by the CLIVAR Working Group on Ocean Model Development where additional fresh water flux into the North Atlantic is specified as a line source around Greenland. Results from that experiment include changes in the salinity distribution in the northern North Atlantic due to the source as well as due to changes in the oceanic circulation. The major response in sea level is found in the North Atlantic, especially in the Nordic Seas and the Arctic Ocean. Most of the sea level signal is due to changes in the large scale oceanic circulation.

Investigating ice shelves using Bayesian Hierarchial Modeling: Preliminary results and comparison to spatial reconstructions of former Arctic Ocean ice shelves

Nina Kirchner

Ice shelves currently exist mainly along the coastline of Antarctica. Over millennia, ice flows from the interior of the continent to the coast to eventually become afloat on the Southern Ocean while still being attached to the inland ice sheet. In contrast, ice shelves are to date virtually non-existing in the Arctic, but the issue of a former Arctic Ocean Ice Shelf is a re-current one among paleoglaciologists [1,2]. Since upon melting, floating ice shelves leave hardly any traces behind, uncertainty regarding their spatio-temporal distribution is high.

To assess the variability in the ice shelf configurations we have previously exploited the BEDMAP data set in a pilot study [3,4]. Note, however, that provision of data on ice shelves is not the primary target in the BEDMAP data set, implying that results obtained in the pilot study are mostly qualitative. In the meantime, high resolution ice shelf-specific data for Antarctic ice shelves has become available e.g. from NASAs ICESat missions [5], and we use this data as input into a Bayesian hierarchial modeling framework.

The basic idea behind a dynamic Bayesian hierarchical modeling (BHM) approach is to break the (statistical) model down into a data part, a process part and a prior part, linking them through the Bayes theorem [6,7]. More specifically, the state of an ice shelf is discretized and represented numerically, e.g., by providing i) length along a central streamline, ii) length at the calving front, iii) ice thickness at the calving front, iv) ice thickness at the grounding line, viz. where the ice becomes afloat, and many others. We write the state vector at instance t as Xt and use simplistic representations of ice flow dynamics to propagate it from time t to t +1 (process part). The state is (almost always) only indirectly observed through a 'measurement operator' H, i.e., Yt = H (Xt) (data part), denoting the observations as Yt. The model is supplemented with prior distributions whose choice is crucial since the system is over-parameterize (prior part). The Bayesian formulation by means of complex Markov chain Monte Carlo simulations allows to infer the state (and other parameters of interests) from the observations. An additional advantage of the methodology is the ability to naturally deliver uncertainty estimates of the parameters.

Due to the complexity of the system, we need present data from Antarctic ice shelves [5,8] to train and constrain the dynamic BHM [9]. In a second step, we will apply the model to Arctic Ice shelves of the Pleistocene. Marine traces can then either be used to further constrain the model or as a model validation tool. Eventually, we will be able to predict possible former Arctic Ocean ice shelf configurations. This approach complements ongoing modeling efforts in numerical ice sheet modeling [10].

Fimbulisen II - modelling: Investigation of the influence of a rift zone on the flow dynamics by means of numerical modelling

Angelika Humbert¹ and Daniel Steinhage²

The definition of zones of similar structure presented in ‘Fimbulisen I – Radar2’ will be used for numerical modelling. The modelling aims to study the viscosity of the distinct zones and to identify the damage related softness of the ice. Former validated simulations of the present-day three-dimensional temperature field of Fimbulisen allows to determine the flow rate factor and thus the temperature dependence precisely. In case the viscosity would solely be determined by the ice temperature, Fimbulisen would be a stiff ice shelf, as it consists of a cold bulk ice mass in the ice stream. The observed flow field on the other hand, points towards a contribution of damage to the viscosity via a stress enhancement factor. The zones of similar structure identified by the classification of satellite radar imagery and airborne radio echo sounding, were incorporated in a 2D diagnostic ice flow model as sub-domains with variable stress enhancement factor and thus treated as zones of different damage related stiffness. The comparison of thousands of simulations with observed velocities at characteristic locations shows that a unique combination of stress enhancement factors exists. Furthermore, the value of the determined enhancement factor in the various regions is linked to the observed damage in the radar data. Finally, we present a synthesis of the radar data and the numerical modelling, which helps to understand the causes for the highly rifted zone in the western part of Fimbulisen.

Evaluation of the criticality of cracks in ice shelves using fracture mechanical concepts

Carolin Plate¹, Ralf Müller¹, Dietmar Gross², Angelika Humbert³ and Matthias Braun⁴

Massive break-up events in ice shelves are often associated with weakened and damaged zones. These zones can comprise pre-existing cracks, which experience a complex stress state due to the ice flow. Using remote sensing data of the Wilkins Ice Shelf a velocity field can be constructed. From the velocity field the stresses are computed by treating the ice as a viscous fluid.

The break up event itself usually occurs on a rather short time scale, thus the elastic response is important for fracture events. Fracture mechanical concepts based on configurational forces in conjunction with finite element simulations are used to evaluate the criticality of representative crack scenarios. By this approach, common hypothesis for ice shelf break up are verified. The scenarios include cracks originating from the top and the bottom of an ice shelf. The stress state of the ice flow is perturbed by the hydrostatic stress state varying through the thickness of the ice shelf. The effect of water in the cracks is taken into account. In a further step the inhomogeneity of the properties through the thickness of the ice shelf will be considered.

Coupling Glimmer-CISM to POP using an Immersed Boundary Method

Asay-Davis Xylar

A proposed method for coupling a dynamical ocean model (The Parallel Ocean Program, POP) to a dynamical ice sheet/ice shelf model (Glimmer-CISM) will be presented. This work aims toward a global, high-resolution simulation of the ocean together with the full Antarctic ice sheet. The ocean model will use an immersed boundary method (IBM) to represent the complex, time-evolving geometry of the ice shelf. The IBM allows for accurate representation of the boundary conditions without the need for a modeling grid that conforms to the boundary or that changes in time.

Ice sheet - ice shelf - ocean interactions: A contribution to ICE2SEA

Silvia Maßmann, Malte Thoma and Klaus Grosfeld

Mass flux from the ice sheet to the ocean is the main cryospheric contributing to sea level rise. In Antarctica, ice sheet and ocean directly interacts via the extended ice shelf regions, where the ice sheet gets afloat. Here, the grounding line represents the transition between the two systems, ice sheet and ice shelf. Its position is highly sensitive to changes in the climate forcing conditions. Ice shelf basal melting influences not only the geometry and dynamics of an ice shelf, but also the mass flux from the inland ice. Investigating the interaction of the three components ice sheet, ice shelf, and ocean requires the simulation of all components in a consistent way. The buttressing effect of ice shelves on the ice sheet is up to date not fully understood and can only be studied with a fully coupled comprehensive approach. For this aim we propose a modelling effort in which we will couple a Full-Stokes ice model, allowing the ice sheet - ice shelf transition to be consistently simulated, with a regional ocean - ice shelf cavity model of high resolution with open boundaries. The first task is to validate the model in physically simple setups comparable with already existing ice sheet models by performing the experiments described in MISMIP (Marine ice sheet model intercomparison project). In a first setup, the position of the grounding line, grounded ice volume and ice thickness at the ice divide are compared for a linear downward-sloping bed, when the model has relaxed to the steady state for given Glen's law parameters. Later experiments will focus on flow law alterations and bedrock-gradient dependencies in order to investigate a possible hysteresis of the ice sheet. These model experiments serve as a basis for future sensitivity estimates and their transfer to other models. In particular, there is an interest in varying boundary conditions like temperature, bedrock slope and the ocean circulation regime to identify their influence on grounding line migration, basal melt rate, ice stream dynamics, ice sheet/ice shelf geometry and ice sheet mass export. The idea is to understand the physical interactions in ice and ocean regimes like they are characteristic for Pine Island Glacier and Filchner-Ronne Ice Shelf.

Coupling of ice-shelf melting and buttressing is a key process in ice-sheets dynamics

Gael Durand

Increase in ice-shelf melting is generally presumed to have triggered recent coastal ice-sheet thinning.

Using a full-Stokes finite element model which includes a proper description of the grounding line dynamics, we investigate the impact of melting below ice shelves. We argue that the influence of ice-shelf melting on the ice-sheet dynamics induces a complex response, and the first naive view that melting inevitably leads to loss of grounded ice is erroneous.

We demonstrate that melting acts directly on the magnitude of the buttressing force by modifying both the area experiencing lateral resistance and the ice-shelf velocity, indicating that the decrease of back stress imposed by the ice-shelf is the prevailing cause of inland dynamical thinning. We further show that feedback from melting and buttressing forces can lead to nontrivial results, as an increase in the average melt rate may lead to inland ice thickening and grounding line advance.

Impact of Antarctic sub-glacial lakes on the ice flow

Malte Thoma^1,2, Klaus Grosfeld² and Christoph Mayer¹
Part I: Andrew M. Smith⁴, John Woodward⁵ and Neil Ross⁶
Part II: Frank Pattyn³

Subglacial lakes are a widespread phenomenon across the Antarctic Ice Sheet, as became clear during the last few years. Most of these lakes are isolated from direct exchange with the atmosphere by several kilometers of ice since millions of years and provide unique environments for potential life forms. The role of the lakes for the subglacial hydrology is still not clear, but there are indications for an active hydrological system underneath the ice sheet, which has an impact on the general ice dynamics. The inaccessibility of this system increases the importance of numerical models to investigate the physical conditions in these environments.

Part I: The unique temperature profile of Subglacial Lake Ellsworth

We apply a three-dimensional fluid-dynamics lake model on the geometry of this lake. We present results, indicating that Subglacial Lake Ellsworth has a unique temperature profile, constituted by the pressure induced by the ice thickness above as well as the slope of the ice-lake interface.

Part II: The impact on the ice flow of Subglacial Lake Vostok

We present model results of a coupled system, consisting of a lake-flow and an ice-sheet model for Subglacial Lake Vostok. The basal mass balance, estimated with a 3D fluid-dynamics lake model is applied as a boundary condition to a Full-Stokes ice-sheet model and its impact on the flow and temperature regime of the ice sheet is analysed. Our results indicate the sensitivity of the ice flow and temperature to adjustments in various boundary conditions and indicate the importance of a Full-Stokes model to such a system.

Airborne observations of the atmospheric boundary layer over the Ronne Polynya

C. Luepkes and J. Hartmann

It is well known that the energy exchange between atmosphere and ocean is especially large over open leads and polynyas. Its quantification by modelling and/or observations is, however difficult, since observations can only be carried out by ship or aircraft with special equipment and the small scale of the involved processes complicates the modelling.

In this talk first results are presented of some flights with the german aircraft Polar 5 over the Ronne polynya during the aircraft campaign JASPER, which has been carried out during February/March 2010 as a joined AWI/BAS campaign. Results document the strong impact of the polynya. It is shown that the transport of heat and momentum over Ronne has a large variability with respect to time and space.

Long term CTD observations under the Ekström Ice Shelf at the PALAOA acoustic observatory

Olaf Boebel

In December 2005, the AWI installed PALAOA (Perennial Acoustic Observatory in the Antarctic Ocean), an energetically autonomic, hydroacoustic listening station on the Ekström ice shelf, eastern Weddell Sea, Antarctica. PALAOA is located close to the ice shelf edge, adjacent to the north-eastern seaward corner of Atka Bay, some 20 km north of the Neumayer station, to which it is linked by wireless communication, allowing real time data access and remote control.

A hydrophone array and a CTD were deployed through 4 boreholes driven by hot-water boring (Nixdorf, 1994) through the approximately 100 m thick floating ice shelf and placed about midway in the water column at a depth of about 155 m, i.e. 70 m below the ice shelf and 90 m above the seafloor. As the boreholes refroze shortly after, the instruments are physically inaccessible. Instruments are permanently connected by cable with the PALAOA container, which hosts the energy system (batteries, fuel cell, solar panels and wind generator), system control (control computer and relay banks), recording (filters, preamplifiers, analogue-digital converters, network attached storage), communication (serial terminal server, wireless local area network bridge), and ancillary (GPS, Webcam) equipment.

CTD data (which is needed in the acoustic context for sound speed estimates) is sampled at half hourly intervals and recorded internally. The provision of power had to be arranged externally from PALAOA, as the open ended nature of this under-ice CTD deployment conflicts with the limited capacity of internal batteries. Data is remotely accessible through a serial interface connected to a terminal server. A script running on a server at Neumayer station collects and backups the data daily and automatically creates plots for system monitoring using an online web interface. Starting in December 2005, the CTD time series now spans more than 4.5 years. Data are continuous with the exception of a 10-month gap between April 2006 and January 2007, which was caused by a breakage in the surface segment of the connecting cable. Temperature and salinity time series clearly depict a strong seasonal cycle. Wintertime temperatures are typically less then -1.8°C, with lowest temperatures reached in early spring. During summer, a pulsed warming is observed, reaching peak temperatures of typically -1.4°C. Salinity reaches its minimum by mid-summer, with values ranging between 33.2 and 33.6. Thereafter, salinity slowly increases towards values of 34.2 which are reached in November. Salinity is then observed to decrease slightly to 34.1, before the next summer minimum is obtained abruptly.

The pressure record is dominated by tidal signatures. Generally, regardless of its direction, any flow past the CTD results in its vertical displacement. At slack water, the CTD hangs straight down at its cable at a maximum depth of 160 m. At regular intervals however, the CTD is significantly displaced upwards, which is indicative of strong oncoming currents. Unfortunately, the pressure data record only provides an uncalibrated measure of current strength and no information on its direction. Overlaid to the pressure record's diurnal and fortnightly tidal patterns, a seasonal cycle exhibits increased upward displacements (hence maximum flow strengths) during summer. This is possibly due to reduced tidal friction or an increased Ekman transport with waning fast ice. These summertime upward CTD displacements appear to increase each year, probably reflecting the slow northward motion of the ice shelf (600 m over the past 4 years), which, coupled with its calving, leads to the CTD being progressively placed into higher velocity regimes of the Antarctic Coastal Current [Núñez-Riboni and Fahrbach, 2009]. Comparison of the pressure record with expected tidal amplitudes from the TPXO7.1 tidal model [Egbert and Erofeeva, 2002] did not immediately reveal an obvious interrelationship, which probably is due to a complex superposition of tidal flow, the Antarctic Coastal Current, deflection effects from the ice shelf and icebergs grounded in Atka Bay, and the wind driven Ekman transport.

This contribution will present the described data and provide a detailed analysis of possible sensor drifts, noting that the instrument is permanently inaccessible for recalibration. Ongoing analysis of the data, focussing on a correlation of the temporal evolution of the CTD data with seasonal ice formation and melting aims at a first interpretation of the data in terms of the underlying driving forces.

The role of ice rises in the stability of an ice sheet: Interpreting radar layers using ice-flow models

D. Docquier, F. Pattyn, K. Matsuoka, D. Callens and H. Conway

Ice rises in coastal areas of Antarctica play a crucial role in the dynamics of the advancing and retreating ice sheet. Most ice rises are islands of grounded ice within the ice shelf, although some of them are still connected to the continental ice sheet. They are characterized by a local ice flow pattern (e.g. Berkner Island) and exert a backpressure on the ice shelf, hence stabilizing the inland ice flow.

In 2008, the BELISSIMA–1 expedition (Belgian Ice Sheet - Shelf Ice Measurements in Antarctica) carried out a number of radio-echo sounding and D-GPS surveys across an ice rise still connected to the main ice sheet in Dronning Maud Land, East Antarctica. The surveyed profiles run across the ice rise itself as well as in the saddle area, connecting the ice rise to the main grounded ice flow. The latter profiles have the advantage that lateral strain components can be neglected (non-convergent ice flow), which facilitates the use of 2D flowline models. Both bedrock elevation and internal layering were identified. Several ice sheet models of different complexity (higher-order) were applied to investigate the nature of internal layer anomalies (such as basal melting), once major effects of surface accumulation and bedrock sensitivity were accounted for.

Effect of numerical resolution on grounding line position in marine ice sheet models

D. Docquier, F.M. Nick, L. Perichon and F. Pattyn

Grounding line migration is a key process affecting the stability of marine ice sheets that rest on a bed lying below sea level (Weertman, 1974; Schoof, 2007). Horizontal shearing (grounded ice sheet) and longitudinal stretching (ice shelf) couple together across a transition zone near the grounding line. Different ice sheet models (SIA0, L1L2, HOM (Pattyn et al., 2006)) are currently being improved in order to accurately model grounding line migration. However, there is a strong dependency of models using a fixed grid on numerical details such as the horizontal grid size (Vieli and Payne, 2005; Durand et al., 2009). We performed a series of experiments with a SIA (shallow-ice approximation) flowline model to identify the most suitable numerical resolution (i.e. grid size) to model grounding line migration.

A two-layer cavity flow model to study interactions beneath ice shelves

Victoria Lee

Observational studies have shown that ice shelves off the coast of West Antarctica in the Amundsen Sea area have experienced rapid thinning during the 1990s (Shepherd et al., 2002; Shepherd et al., 2004). In particular, it is believed that oceanographic changes in the Amundsen Sea are responsible for the thinning and acceleration of Pine Island Glacier. An earlier study looking at melt rates generated by a plume model beneath the ice shelf in Pine Island Bay suggested that there was a strong feedback mechanism between the plume and topography of the ice shelf underside near the grounding line (Payne et al., 2007). The influence of ocean properties on the melt rates could not, however, be fully investigated by the simple plume model because it assumes that the ambient ocean is stationary and whose properties do not vary horizontally.

We have modified the above 2D plume model to consist of two-layers; the upper one represents the buoyant, melt-water rich plume and the other represents the circulating ambient ocean water. The addition of an active ambient layer to the 2D model allows ocean properties at the ice shelf front to advect around the cavity. This new model will provide spatial patterns of basal melt rates that will include both the effects of the geometry of the cavity beneath an ice shelf and changes in ocean properties at the shelf's front.

We have developed a simple 2D model so that it can be coupled to an ice sheet model. Momentum of the cavity flow is governed by geostrophic balance modified by a linear drag law and is solved using the rigid lid streamfunction method. Interactions between the layers are parameterised through the entrainment rate, which depends on the relative velocities, thicknesses and densities of the layers and can be positive or negative.

The model will be applied to the Pine Island Ice Shelf. Results from early simulations with realistic topography of the underside of the the shelf and a flat bedrock using parameter values used by Payne et al. (2007) suggested that melt rates are concentrated close to the grounding line where the slope of the underside of the shelf is greatest and they are around 50 m/yr.

Thermodynamics of brine infiltration in ice shelves

Martin O'Leary

Since the 1960s it has been known that large quantities of brine exist within several Antarctic ice shelves, including those which have been identified as being vulnerable to climate-induced retreat and collapse. This brine has been hypothesized as possibly leading to enhanced calving through hydrofracture mechanisms. However, there has been very little study of the mechanisms through which this brine layer is formed, and no predictive model exists for the location and properties of the brine.

We present a model for the development of such a brine layer, incorporating the interaction between thermodynamic considerations and the permeability of the firn. We examine the behaviour of this model, and show that it is consistent with observed characteristics of known brine intrusions.

We apply the model to investigate whether such brine intrusions could be responsible for regular calving. Using linear elastic fracture mechanics, we study the hydrofracture potential of the brine layer, and determine calving rates due to the brine infiltration.

RTOPO-1: A consistent data set for Antarctic ice shelf topography and global ocean bathymetry

Ralph Timmermann et al.

Sub-ice shelf circulation and freezing/melting rates depend critically on an accurate and consistent representation of cavity geometry (i.e. ice-shelf draft and ocean bathymetry). Existing global or pan-Antarctic data sets have turned out to contain various inconsistencies and inaccuracies. The goal of this work is to compile independent regional fields into a global data set. We use the S-2004 global 1-minute bathymetry as the backbone and add an improved version of the BEDMAP topography for an area that roughly coincides with the Antarctic continental shelf. Locations of the merging line have been carefully adjusted in order to get the best out of each data set. High resolution gridded data for the Amery, Fimbul, Filchner-Ronne, Larsen C and George VI Ice Shelves and for Pine Island Glacier have been carefully merged into the ambient ice and ocean topographies. Multibeam ship survey data for bathymetry in the former Larsen B cavity and the southeastern Bellingshausen Sea have been obtained from the data centers of Alfred Wegener Institute (AWI), British Antarctic Survey (BAS) and Lamont-Doherty Earth Observatory (LDEO), gridded, and again carefully merged into the existing bathymetry map. The resulting global 1-minute data set contains consistent masks for open ocean, grounded ice, floating ice, and bare land surface.